FET-ENVIS Final Report

• Main Authors: Sandven, Stein, Hamre, Torill, Bogdanov, Andrej, Steinhage, Axel, Winkel, Carsten, Aplin, Paul, Vieira, Carlos, Mather, Paul
• Format: Report
• Language: English
• Published: 2001
• Subjects: Environment; Information; Extraction; Marine; Terrestrial; Sea ice; Pollution; Agriculture; Human relief; Earth Observation; Detection; Classification
• Online Access: https://zenodo.org/record/7638621; https://doi.org/10.5281/zenodo.7638621

The overall objective of this project has been to investigate methods for efficient extraction and synthesis of environmental information using multi-source data analysis methods, which have the potential for reuse in a wide range of environmental problems. Four complementary case studies were chosen from the partners' field of expertise. The selected case studies covered both applications from the terrestrial and marine domain, to ensure a broad basis for testing cross-domain application of algorithms and enabling extrapolation of generalized results from the study as a whole. The case studies investigated problems related to sea ice monitoring at high latitudes, detection of marine pollution, agricultural crop classification and use of very high-resolution satellite imagery for support of human relief operations. A review of state-of-the-art was conducted for all case studies, and the partners has implemented and assessed multi-source data analysis algorithms for the three first application domains. In humanitarian relief operations EO-data have only recently started to be used. Analysis methods are therefore mainly based on interpretation by operators and automated methods are not used. However, as this application becomes more mature, it can potentially adopt automated methods from the three other application domains. The overall conclusions from the project are as follows: • Partners have developed and assessed a set of multi-source data analysis algorithm for selected case studies, and in addition successfully applied two of these algorithms in another domain than for which it was originally implemented. The investigated algorithms represent state-of-the-art in the respective application domains. • Development of cross-domain algorithms for extraction of environmental parameters on a wider basis has been motivated by the identification of three basic parameters that describe the nature of an environmental problem. Such a categorization scheme can enable an EIS to make a best match between candidate ...